From: Howard C. Berkowitz (hcb@gettcomm.com)
Date: Fri Apr 23 2004 - 17:24:29 GMT-3
At 2:29 PM -0400 4/23/04, BALAKRISHNAN Balaji wrote:
>Hi group,
>
>I have a clarification on RFC1812 about route forwarding decision..
>
>When a router received a packet,
Maybe I'm missing something fundamental in your question. Yes, the
router receives the packet because it's in promiscuous mode.
The "local vs. remote" assumption means that if a packet destination
resolves to be on the same subnet as the source, then the destination
is reachable from the source through layer 2 connectivity. This is
obviously the case on point-to-point and broadcast media, and is the
reason that NBMA can be so problematic -- NBMA violates the local vs.
remote assumption.
We are talking about a single broadcast medium such as an Ethernet,
correct? The source host should ARP, and, if the destination is on
the same subnet, it should respond with its MAC address. The source
will subsequently send unicast packets to that address.
How does the router get involved if this is what you are describing?
>if the destination ip is part of its
>directly connected interface subnet , it will do the arp request for
>that destination ip out to that interface. Now, we can change that
>behavior by configuring more specific route for that destination ( say
>/32 host route) pointing to different gateway. On this case, router
>forwards the packet to the next-hop defined. right ???
>
>But my manager don't agree with that. He says if destination IP address
>in the packet is on the same subnet as any Ethernet interface of the
>router, then the router should not do any route lookup at all, and must
>immediately forward the packet on that matching interface . Ie. more
>specific route present in the router's routing table is then not
>relevant. He pointing to the RFC 1812 section "5.2.4.2 Local/Remote
>decision ". The section goes line this ..the following algorithm MUST be
>used to determine if the Immediate Destination is directly accessible
>quote
> * isolate the network prefix used by the interface.
> * Isolate the corresponding set of bits from the IP Destination
>Address of the packet.
> * Compare the resulting network prefixes. If they are equal to each
>other, the packet can be transmitted through the corresponding network
> interface.
> * If the destination was not a member of a directly connected network
>prefix, the IP Destination is accessible only through some other
> router.
>unquote
>
>I believe it does not consider all the possibilities for a route
>decision. Do anybody know better RFC reference that I can point to my
>manager to prove that I am right ???
>
>- bala.
>
>_______________________________________________________________________
>Please help support GroupStudy by purchasing your study materials from:
>http://shop.groupstudy.com
>
>Subscription information may be found at:
>http://www.groupstudy.com/list/CCIELab.html
This archive was generated by hypermail 2.1.4 : Mon May 03 2004 - 19:48:54 GMT-3